Platform Events Guide

Platform Events Guide

Overview

Platform Events enable event-driven architecture in Salesforce. They provide a scalable, asynchronous messaging system for real-time integrations.

Event Types

Standard Volume

• Up to ~2,000 events per hour
• Included in all Salesforce editions
• Standard delivery guarantees

High Volume

• Millions of events per day
• At-least-once delivery
• 24-hour retention for replay
• May require additional entitlement

When to Use Platform Events

Creating Platform Events

Via Metadata (Recommended)

<?xml version="1.0" encoding="UTF-8"?>
<CustomObject xmlns="http://soap.sforce.com/2006/04/metadata">
    <deploymentStatus>Deployed</deploymentStatus>
    <eventType>HighVolume</eventType>
    <label>Order Update Event</label>
    <pluralLabel>Order Update Events</pluralLabel>
    <publishBehavior>PublishAfterCommit</publishBehavior>
    <fields>
        <fullName>Order_Id__c</fullName>
        <label>Order ID</label>
        <type>Text</type>
        <length>18</length>
    </fields>
    <fields>
        <fullName>Status__c</fullName>
        <label>Status</label>
        <type>Text</type>
        <length>50</length>
    </fields>
</CustomObject>

File Location

force-app/main/default/objects/Order_Update_Event__e/Order_Update_Event__e.object-meta.xml

Publishing Events

From Apex

// Single event
Order_Update_Event__e event = new Order_Update_Event__e();
event.Order_Id__c = orderId;
event.Status__c = 'Shipped';

Database.SaveResult result = EventBus.publish(event);
if (result.isSuccess()) {
    System.debug('Event published: ' + result.getId());
}

// Multiple events
List<Order_Update_Event__e> events = new List<Order_Update_Event__e>();
// ... populate events
List<Database.SaveResult> results = EventBus.publish(events);

From Flow

1. Create Record element
2. Select Platform Event object
3. Map field values

From Process Builder

1. Add Immediate Action
2. Select "Create a Record"
3. Choose Platform Event

Subscribing to Events

Apex Trigger

trigger OrderUpdateSubscriber on Order_Update_Event__e (after insert) {
    for (Order_Update_Event__e event : Trigger.new) {
        System.debug('Order ' + event.Order_Id__c + ' is now ' + event.Status__c);
        // Process event
    }

    // Set checkpoint for durability
    EventBus.TriggerContext.currentContext().setResumeCheckpoint(
        Trigger.new[Trigger.new.size() - 1].ReplayId
    );
}

Flow (Record-Triggered)

1. Create Platform Event-Triggered Flow
2. Select Platform Event object
3. Build logic with event data

External (CometD)

External systems can subscribe using CometD streaming:

/event/Order_Update_Event__e

Publish Behavior

PublishAfterCommit (Default)

• Event published after transaction commits
• If transaction rolls back, event NOT published
• Recommended for most cases

PublishImmediately

• Event published immediately
• Event still published even if transaction rolls back
• Use when external system must be notified regardless of outcome

Durability & Replay

Replay ID

Each event has a unique ReplayId for tracking and replay:

String replayId = event.ReplayId;

Resume Checkpoint

Set checkpoint to ensure durability:

// In trigger
EventBus.TriggerContext.currentContext().setResumeCheckpoint(lastReplayId);

If trigger fails after checkpoint, processing resumes from that point.

Retention

• High Volume events: 24 hours
• Standard Volume: 24 hours

Best Practices

Publishing

1. Batch events when publishing multiple
2. Check SaveResults for publish failures
3. Use meaningful correlation IDs for tracking
4. Include timestamp for ordering
5. Keep payloads small - use IDs, not full records

Subscribing

1. Always set resume checkpoint in triggers
2. Don't throw exceptions - catch and log errors
3. Process idempotently - events may replay
4. Keep processing lightweight - queue heavy work
5. Handle duplicates using correlation ID

Design

1. Event granularity - not too fine, not too coarse
2. Include enough context but not entire records
3. Version your events if schema evolves
4. Document event contracts for consumers

Error Handling

Publish Errors

List<Database.SaveResult> results = EventBus.publish(events);
for (Integer i = 0; i < results.size(); i++) {
    if (!results[i].isSuccess()) {
        for (Database.Error err : results[i].getErrors()) {
            System.debug('Publish failed: ' + err.getMessage());
        }
    }
}

Subscriber Errors

trigger MySubscriber on My_Event__e (after insert) {
    for (My_Event__e event : Trigger.new) {
        try {
            processEvent(event);
        } catch (Exception e) {
            // Log error, don't throw
            System.debug('Error processing ' + event.ReplayId + ': ' + e.getMessage());
            // Create error log record
        }
    }

    // Still set checkpoint even if some failed
    EventBus.TriggerContext.currentContext().setResumeCheckpoint(lastReplayId);
}

Monitoring

Setup → Platform Events

• View event definitions
• Check usage metrics
• Monitor delivery status

Event Delivery Failures

Check for:

• Unhandled exceptions in triggers
• Apex CPU timeout
• Governor limit errors

Event Publishing

Query EventBusSubscriber for subscription health:

SELECT Id, Position, ExternalId, Name, Status, Tip
FROM EventBusSubscriber
WHERE Topic = 'Order_Update_Event__e'

Limits

External Integration

Subscribe from External System

Use CometD client to connect to Streaming API:

Endpoint: /cometd/62.0
Channel: /event/Order_Update_Event__e

Publish from External System

Use REST API:

POST /services/data/v62.0/sobjects/Order_Update_Event__e
Content-Type: application/json

{
    "Order_Id__c": "001xx000003NGSFAA4",
    "Status__c": "Shipped"
}